AdaMem: Adaptive User-Centric Memory for Long-Horizon Dialogue Agents

2026-03-17

AI Summary

Existing dialogue memory systems often suffer from excessive semantic dependency, fragmented storage, lack of temporal-causal coherence, and static memory granularity, hindering their ability to support long-term personalized interaction. To address these limitations, this work proposes a unified, adaptive user-centric memory framework that integrates four complementary modules—working memory, episodic memory, persona memory, and graph memory—within a multi-level organizational structure. The framework employs query-conditioned dynamic retrieval routing to enable context fusion that is both semantically and relationally aware, while adaptively adjusting memory granularity based on interaction demands. Evaluated on the LoCoMo and PERSONAMEM benchmarks, the proposed approach achieves state-of-the-art performance, substantially enhancing user modeling and reasoning capabilities in long-horizon dialogues.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Natural Language Processing: Conversational AI/Dialog Systems
• Cognitive Modeling & Cognitive Systems: Adaptive Behavior

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Search and Retrieval-Augmented AI: Personalized, context-aware and across-device search

Abstract

Large language model (LLM) agents increasingly rely on external memory to support long-horizon interaction, personalized assistance, and multi-step reasoning. However, existing memory systems still face three core challenges: they often rely too heavily on semantic similarity, which can miss evidence crucial for user-centric understanding; they frequently store related experiences as isolated fragments, weakening temporal and causal coherence; and they typically use static memory granularities that do not adapt well to the requirements of different questions. We propose AdaMem, an adaptive user-centric memory framework for long-horizon dialogue agents. AdaMem organizes dialogue history into working, episodic, persona, and graph memories, enabling the system to preserve recent context, structured long-term experiences, stable user traits, and relation-aware connections within a unified framework. At inference time, AdaMem first resolves the target participant, then builds a question-conditioned retrieval route that combines semantic retrieval with relation-aware graph expansion only when needed, and finally produces the answer through a role-specialized pipeline for evidence synthesis and response generation. We evaluate AdaMem on the LoCoMo and PERSONAMEM benchmarks for long-horizon reasoning and user modeling. Experimental results show that AdaMem achieves state-of-the-art performance on both benchmarks. The code will be released upon acceptance.

Problem

• long-horizon dialogue
• user-centric memory
• memory coherence
• adaptive granularity
• external memory

Innovation

• adaptive memory
• user-centric dialogue
• graph-based retrieval
• long-horizon reasoning
• memory organization